DESCRIPTION: The goal in this application is to integrate standardized medical vocabularies and record structures into an EMRS that supports clinicians' access to medical information regardless of site and database accessed. The proposed EMRS is expected to (a) display and process information for specific clinical settings,(b) communicate information between multiple clinical sites, (c) retrieve information and display data consistently independent of the underlying database, and (d) automatically map from local to standardized medical vocabularies. The project will be carried out at three sites, Children's Hospital in Boston, the MIT Lab for Computer Science, and the Albert Einstein College of Medicine. The project will address a number of problems encountered with the emerging EMRS. As medical vocabularies and EMRS structures become standardized, hundreds of existing clinical databases may not conform to these standard formats. The proposed system would function independent of both clinical site and database, and thus, provide for use of many databases. While standards for interchanging data have progressed, the development and use of standardized medical vocabularies has been slow. The electronic medical record systems developed to date perform well at the development sites, but have been less than satisfactory when transported to other sites. Typically, patient records are spread among multiple databases, each of which may be based on a different technology, have come from a different vendor, and use disparate interfaces and query languages. The need is for a client-server architecture for heterogeneous database access such as SIMS. SIMS is a knowledge base system supporting a semantic model of the problem domain, and uses this model to reformulate uniform queries as database-specific queries. In SIMS, associate databases are treated as information "servers", whereas the semantic model and the query reformulation methods form the "client".